Use case · NPS deep dive

NPS is a number. The verbatims are the story.

Your NPS is +32. Now what?

Most teams stop at the score. They track it, report it, and feel good or bad about it. But the number alone tells you nothing about what to do next.

The answers are in the verbatims — the open-text responses where customers explain their score. “The product is great but the support is terrible.” “I’d recommend it if the reporting were better.” “Switched to a competitor last month.”

These verbatims are the richest data in your NPS programme. And almost nobody analyses them at depth.

InsightNarrator runs five analysis techniques on every NPS dataset, automatically:

  • ABSAWhich aspects of the experience drive promoter and detractor scores.
  • Sentiment Driver AnalysisWhat changes the score.
  • Feedback ClassificationHow many “detractor” comments are actually feature requests.
  • Pain Point DiscoveryThe friction points that prevent passives from becoming promoters.
  • Loyalty Driver AnalysisWhy promoters advocate.

NPS overview

4,200 responses

+32

48% promoters36% passives16% detractors
  • Support response time−0.71
    Detractor driver
  • Reporting limitations23%
    Passive blocker
  • “It just works”71%
    Promoter theme

The NPS trap

Why a number without context is dangerous

An NPS of +42 looks good on a dashboard. But here’s what the number hides.

+42 overall, but −8 in your largest enterprise segment.

The aggregate masks a retention crisis.

+42 this quarter, up from +38.

But you changed the survey timing and the distribution channel. The improvement isn’t real.

+42 with 60% passives.

Your customers aren’t detractors, but they’re not advocates either. They’re one competitor pitch away from switching.

+42 with 90% of promoters giving no verbatim reason.

You don’t know why they like you. How do you protect what you can’t identify?

The score tells you where you stand. The verbatims tell you why — and what to do about it.

“We spent two years trying to move our NPS from +28 to +35. We hit +35 the quarter after we started analysing verbatims with InsightNarrator. Not because the tool changed the score — it changed what we fixed.”

CX Director, B2B SaaS (composite voice)

The five analyses

Five techniques, run automatically on every NPS dataset

Analysis 1

ABSA — which aspects drive your score

ABSA breaks down the NPS verbatims into specific aspects of the experience and scores sentiment per aspect. Instead of “overall, customers are moderately positive,” you get an aspect-level split between promoters and detractors.

What this tells you: Customer Support has the largest promoter–detractor gap (+1.05). Fix support, and you move more detractors than any other intervention.
The aspects come from your governed taxonomy — so “Customer Support” means the same thing this quarter and next quarter. The comparison is valid.
AspectPromotersDetractorsDelta
Product Quality+0.81+0.22+0.59
Customer Support+0.34−0.71+1.05
Onboarding+0.52−0.38+0.90
Pricing+0.12−0.55+0.67
Reporting+0.45−0.15+0.60

Analysis 2

Sentiment Driver Analysis — what causes the score to change

Driver analysis identifies the variables that most strongly correlate with NPS shifts, ranking each driver by influence on score variance.

What this tells you: Support response time is the single biggest driver. Improve it by one hour on average and the NPS model predicts a +4.2 point lift. Number of features used is second — customers who adopt 3+ features score 18 points higher. (Illustrative output — actual lifts vary by dataset.)
This is the difference between guessing what to fix and knowing.
  1. 1. Support response time34%
  2. 2. Number of features used22%
  3. 3. Time since last interaction15%
  4. 4. Billing clarity11%
  5. 5. Onboarding completion8%

Share of NPS score variance explained.

Analysis 3

Feedback Classification — what detractors are really saying

Classification sorts every verbatim into categories, split by NPS band.

What this tells you: 41 detractors are actually telling you about missing features — these are product opportunities, not satisfaction problems. 28 detractors show churn signals — these need immediate intervention. The “detractor” label doesn’t mean the same thing for all of them.
CategoryPromotersPassivesDetractors
Praise (Product)340452
Feature Request286741
Complaint (Support)52289
Complaint (Pricing)31834
Churn Signal0428
Comparison (Competitor)2819
Question12153

Analysis 4

Pain Point Discovery — what prevents promoters

Pain point analysis surfaces the friction hiding in passive verbatims.

What this tells you: Passives aren’t passive — they’re waiting for specific improvements. Address reporting, mobile, and integrations, and a significant portion become promoters.

Reporting limitations

23% of passives

I can’t get the data I need without exporting to Excel.

The #1 reason passives don’t become promoters.

Mobile experience

15% of passives

The mobile app is slow and missing features.

Integration gaps

11% of passives

Doesn’t connect to our CRM.

Analysis 5

Loyalty Driver Analysis — why promoters recommend you

Loyalty driver analysis isolates the factors that create genuine advocacy, not just satisfaction.

What this tells you: Reliability and time savings are your advocacy engine. Don’t compromise these for new features. Every roadmap decision should be evaluated against: “Does this maintain reliability and time savings?”
  1. 1. Product reliability71%

    It just works.

  2. 2. Time savings58%

    Saves me 3 hours per week.

  3. 3. Support quality44%

    Support team actually solves problems.

  4. 4. Team collaboration31%

    My whole team uses it.

Share of promoters mentioning each theme.

Worked example

A B2B SaaS company, 12 months, real numbers

  1. Starting point

    NPS +28. 40% detractors. No systematic verbatim analysis.

  2. Month 1

    Uploaded 4,200 NPS responses from the past 18 months. Ran all five analyses.

  3. Month 2

    Based on ABSA and driver analysis, invested in support response time (average dropped from 14 hours to 4 hours).

  4. Month 4

    Based on pain point discovery, launched a reporting redesign (the #1 passive blocker).

  5. Month 6

    Based on classification, routed 41 “feature request” detractors to the product team. 12 became roadmap items.

  6. Month 9

    NPS reached +41. Detractor rate dropped from 40% to 24%.

  7. Month 12

    NPS stabilised at +44. Promoter rate hit 61%. Churn down 3.2 percentage points.

At a glance

NPS
+28 → +44
Detractors
40% → 24%
Promoters
→ 61%
Churn
−3.2 pts

The score improved because they stopped optimising the number and started fixing what the verbatims revealed.

Comparison

Why this is different from NPS dashboards

Traditional NPS toolInsightNarrator
Tracks the scoreTracks the score and explains it
Manual verbatim taggingAutomated, governed classification
“Detractor” is a label“Detractor” is a starting point for 5 analyses
Theme tracking is manual and inconsistentABSA uses your taxonomy — consistent across survey waves
No connection to business riskC-Risk identifies which detractors are about to escalate
Reports every quarterAnalyse on demand, as often as you import data

Your NPS verbatims are sitting on a hard drive, untapped.

Upload your last NPS export and get your first deep analysis in minutes. No setup, no configuration — Scoty reviews your data and recommends where to start.

EU/Swiss hosting option — your data stays yours.

EU/Swiss hostingGDPR-readyNo credit card required